English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Let us rethink the real-world scenarios that require human motion predictiontechniques, such as human-robot collaboration. Current works simplify the taskof predicting human motions into a one-off process of forecasting a shortfuture sequence (usually no longer than 1 second) based on a historicalobserved one. However, such simplification may fail to meet practical needs dueto the neglect of the fact that motion prediction in real applications is notan isolated ``observe then predict'' unit, but a consecutive process composedof many rounds of such unit, semi-overlapped along the entire sequence. As timegoes on, the predicted part of previous round has its corresponding groundtruth observable in the new round, but their deviation in-between is neitherexploited nor able to be captured by existing isolated learning fashion. Inthis paper, we propose DeFeeNet, a simple yet effective network that can beadded on existing one-off prediction models to realize deviation perception andfeedback when applied to consecutive motion prediction task. At each predictionround, the deviation generated by previous unit is first encoded by ourDeFeeNet, and then incorporated into the existing predictor to enable adeviation-aware prediction manner, which, for the first time, allows forinformation transmit across adjacent prediction units. We design two versionsof DeFeeNet as MLP-based and GRU-based, respectively. On Human3.6M and morecomplicated BABEL, experimental results indicate that our proposed networkimproves consecutive human motion prediction performance regardless of thebasic model.

DeFeeNet: Consecutive 3D Human Motion Prediction with Deviation Feedback